spacytextblob 0.1.1

A TextBlob sentiment analysis pipeline compponent for spaCy.

spaCyTextBlob

A TextBlob sentiment analysis pipeline compponent for spaCy.

Table of Contents

• Install
• Usage
  • How to load the package in spaCy pipeline
  • How to use the pipeline
• API
• Reference and Attribution

Install

Install spaCyTextBlob from pypi.

pip install spacytextblob

TextBlob also requires some data to be downloaded before getting started.

python -m textblob.download_corpora

Usage

How to load the package in spaCy pipeline

import spacy
from spacytextblob.textblob import SpacyTextBlob

nlp = spacy.load('en_core_web_sm')
spacy_text_blob = SpacyTextBlob()
nlp.add_pipe(spacy_text_blob)

# pipeline contains component name
print(nlp.pipe_names) 

['tagger', 'parser', 'ner', 'spaCyTextBlob']

How to use the pipeline

By adding SpacyTextBlob into the pipeline sentiment analysis is perofmed on the doc everytime you call nlp.

text = "I had a really horrible day. It was the worst day ever!"
doc = nlp(text)
print('Polarity:', doc._.polarity)
print('Sujectivity:', doc._.subjectivity)
print('Assessments:', doc._.assessments)

Polarity: -1.0
Sujectivity: 1.0
Assessments: [(['really', 'horrible'], -1.0, 1.0, None), (['worst', '!'], -1.0, 1.0, None)]

text = "Wow I had just the best day ever today!"
doc = nlp(text)
print('Polarity:', doc._.polarity)
print('Sujectivity:', doc._.subjectivity)
print('Assessments:', doc._.assessments)

Polarity: 0.55
Sujectivity: 0.65
Assessments: [(['wow'], 0.1, 1.0, None), (['best', '!'], 1.0, 0.3, None)]

API

To make the usage simpler spacy provides custom extensions which a library can use. This makes it easier for the user to get the desired data. The below tables summaries the extensions.

spacy.Doc extensions

Extension	Type	Description	Default
doc._.polarity	Float	The polarity of the document. The polarity score is a float within the range [-1.0, 1.0].	None
doc._.sujectivity	Float	The subjectivity of the document. The subjectivity is a float within the range [0.0, 1.0] where 0.0 is very objective and 1.0 is very subjective.	None
doc._.assessments	tuple	Return a tuple of form (polarity, subjectivity, assessments ) where polarity is a float within the range [-1.0, 1.0], subjectivity is a float within the range [0.0, 1.0] where 0.0 is very objective and 1.0 is very subjective, and assessments is a list of polarity and subjectivity scores for the assessed tokens.	None

Reference and Attribution

• TextBlob
  • https://github.com/sloria/TextBlob
  • https://textblob.readthedocs.io/en/latest/
• negspaCy (for inpiration in writing pipeline and organizing repo)
  • https://github.com/jenojp/negspacy
• spaCy custom components
  • https://spacy.io/usage/processing-pipelines#custom-components